Wysiwyg parser escaping markup

> Hi,
> I added code to the Wysiwyg editor to escape markup within text nodes. To illustrate here is an example from the (javascript) test-suite:
> it("should escaped text containing link markup", function() {
> expect("To create a link, put it between <code>*</code>").shouldParseInto("To create a link, put it between ==\\*==");
> });
>
> This works fine, the problem arises with:
> it("should escaped text containing code markup", function() {
> expect("To make something <code>monospaced</code>, surround it with <code>==</code>").shouldParseInto("To make something ==monospaced==, surround it with ==\\=\\===");
> });
> Should text containing '==' translate into '\==' OR '\=\='.
> rendering ' ==\==== ' into html [1] results in:
>     '<p> <code>=</code>= </p>'
> whereas rendering  '==\=\=== ' into html results in:     '<p> <code>==</code> </p>'
>
> The html reflects the PRDocument structure. It seems that PRDocumentParser doesn't parse ' ==\==== ' as expected.Worse the sequence: wiki text-> PRDocument -> wiki text fails:
>      PRWikiWriter write: (PRDocumentParser parse:  ' ==\=\=== ')  => ' ==\==== '
> However I have a fix. I've locally modified PRDocumentParser class>>#escape:using: so that any multi-character markup is multiply escaped:
>  '@@' -> '\@\@' '--' => '\-\-'
> This seems to solve the ambiguous parsing problem and the round-trip problem. Great. However many tests will need rewriting to reflect that any multi-character markup is escaped with multiple slashes. Before fixing the tests, I wanted to check that people thought it made sense.
>
> Thanks
> Nick
> [1] |  document renderer |
> document := PRDocumentParser parse: ' ==\==== '. renderer := PRViewRenderer new.
> ^ WAHtmlCanvas builder fullDocument: false;
> render: [ :r | renderer withinContentDo: [ renderer start:  document in: self on: r ] ].
>

> I've checked in the changes. There were only a couple of tests that needed modifying, they acted as base tests for derived classes and made the number of test failures seem large. If I've missed something I'm happy to roll-back:
>
> Name: Pier-Model-NickAger.411Author: NickAgerTime: 18 July 2011, 10:47:33 pmUUID: d23e8ad0-6e18-4f69-a7fc-7399a8493379Ancestors: Pier-Model-NickAger.410
>
> modified the way markup is escaped, so that:
> '@@' is now escaped as: '\@\@' rather than as previously: '\@@'
> Name: Pier-Tests-Model-NickAger.23
> Author: NickAgerTime: 18 July 2011, 10:44:42 pmUUID: d6cb4015-84c6-45c1-8daa-6f3357f63d92Ancestors: Pier-Tests-Model-NickAger.22, Pier-Tests-Model-lr.21
> modified tests to reflect changes in escaped markup:
>
> '@@'  => '\@\@'
>
>

> Hi,
>
> Thank, I will look at it when I am back home. My only worry was that
> if you had a single = or @ in the text, it would be escaped.
>
> No that shouldn't happen as the only text to be escaped are:
>
> PRDocumentParser buildTextMatcher keys =>  #('__' '''''' '==' '*' '{{{' '--' '^^' '@@' '+' '""')
>
>
> PRWikiWriter write: (PRDocumentParser parse:  ' my address is [hidden email] ') => ' my address is [hidden email] '
>
> All it means is that if you want to escape '@@' you can escape it as '\@\@\ or '\@@', but it escaper will now escape it as '\@\@' so:
>
> PRWikiWriter write: (PRDocumentParser parse:  ' annotations start with: \@@ ') =>  ' annotations start with: \@\@ '
> and:
>
> PRWikiWriter write: (PRDocumentParser parse:  ' annotations start with: \@\@ ') =>  ' annotations start with: \@\@ '
> Nick
>